Using GIS in Global Health
Using GIS in Global Health
Using GIS in Global Health
Information Systems
to Meet Global Health
Challenges
Marc Cunningham
SR-15-126
December 2015
Using Geographic
Information Systems
to Meet Global Health
Challenges
Marc Cunningham
MEASURE Evaluation is funded by the U.S. Agency for International Development (USAID)
under cooperative agreement AID-OAA-L-14-00004 and is implemented by the Carolina
Population Center at the University of North Carolina at Chapel Hill, in association with
ICF International; John Snow, Inc.; Management Sciences for Health; Palladium; and Tulane
University. The opinions expressed in this publication do not necessarily reect the views of
USAID or the United States government.
December 2015
SR-15-126
Acknowledgments
This document was informed by discussions at meetings of the MEASURE Evaluation GIS Technical
Working Group (2015) and the Maternal Mortality Reporting and Mapping working group (also in
2015). The author, Marc Cunningham, is a GIS advisor and data analystat John Snow, Inc., one of
MEASURE Evaluations implementing partners. He prepared this report with support from John
Spencer, senior GIS technical specialist at the Carolina Population Center at the University of North
Carolina at Chapel Hill.
Cover photograph by MEASURE Evaluation, showing HIV program managers in Iringa Region,
Tanzania, using GIS maps to allocate care and treatment services to areas of greatest need.
ii
Table of Contents
Acknowledgments ......................................................................................................ii
Abbreviations ...........................................................................................................iv
Abstract ....................................................................................................................v
GIS and Global Development Priorities ........................................................................ 1
Illustrative Examples .................................................................................................. 2
Achieving an AIDS-free generation ...............................................................................2
Use of GIS to support the scale-up of HIV prevention and treatment activities in
Tanzania ............................................................................................................ 3
Protecting communities from infectious diseases .............................................................5
Use of GIS to coordinate responses to the Ebola outbreak in West Africa ................. 6
Ending preventable child and maternal deaths ..............................................................8
Use of GIS to identify gaps in comprehensive emergency obstetric care coverage in
Uganda ............................................................................................................ 8
Mapping for microplanning: reaching every village ............................................... 9
GISs added value ................................................................................................11
The Current GIS Environment: Increases in Availability of Geospatial Tools, Human
Resources, and Data ................................................................................................ 12
Geospatial tools .................................................................................................. 12
Human resources ................................................................................................. 12
Data ................................................................................................................... 13
Conclusion ............................................................................................................. 13
References ............................................................................................................. 15
Appendix: MEASURE Evaluation Geospatial Resources ................................................ 17
iii
Abbreviations
BEmONC
CEmONC
GIS
IDSR
M&E
QGIS
RHIS
USAID
VMMC
WHO
iv
Abstract
Recent years have seen tremendous growth in interest in geographic information system (GIS) technology.
These systems manage data, facilitate analysis, and generate effective information products that can
support decision making. Considerable investment in health systems has increased the availability of
reliable data, making GIS well suited to support global health and development activities.
At its most effective, GIS is a tool employed in pursuit of a larger objective, such as improving response to
disease outbreaks, increasing access to treatment, or reducing maternal mortality. It enhances the ability
of program managers to distribute services efficiently and equitably. As a supporting tool, its value may be
overlooked if attention is limited to service delivery and health outcomes.
This document presents specific examples of how GIS has served programs associated with key global
health and development priorities. It is intended for program managers, technical specialists, and decision
makers.
The opportunities for GIS for health continue to grow rapidly, due in part to three fundamental changes
over the past decade: increases in data, improvements in software, and growing capacity.
High-quality spatially referenced data, including both health facility locations and administrative
boundaries, are increasingly available for general use. This may reflect increased demand for
geospatial analyses and information products, as well as the broader open data movement.
Classic proprietary systems (e.g., ArcGIS Online and ERDAS IMAGINE) have gotten better
and more user-friendly; there is a growing number of open-source software options with high
GIS capacity (e.g., QGIS) or more limited mapping capacity (e.g., Google Earth). Web mapping
options (e.g., ArcGIS online, Google Fusion Tables, and Drupal) have seen substantial growth
as a viable method for sharing large data sets with key constituencies. There has been a blending
between GIS and data management tools (DHIS 2 and EpiInfo), general data analysis tools (R
and STATA), and data visualization (Drupal).
More people with basic geospatial skills are available for employment, and more resources to train
them are available, as well. This human resources pool includes the historical geospatial
professions (GIS analysts, geographers, and cartographers, etc.); monitoring and evaluation
professionals familiar with health programs and health statistics; as well as a growing number
from the data science field, with its focus on database management, analysis of big data, and data
visualization.
Illustrative Examples
The following examples show how improved access to geographic, granular information has supported
health programs seeking to reach major global health goals: the creation of an AIDS-free generation,
ending preventable child and maternal deaths, and protecting communities from infectious diseases.
Achieving an AIDS-free generation
At the national or global level, great strides have been made in HIV treatment and prevention. National
declines, however, can mask subnational clusters experiencing gaps in service, or can mask specific
subpopulations where HIV prevalence and transmission remain high.
as female sex workers, orphans and vulnerable children, men who have sex with men, and people who
inject drugs, must be identified for improved outreach and consequent service delivery while maintaining
confidentiality to ensure protection of rights and privacy.
Maps can guide HIV program implementation, by supporting program managers efforts to assess
geographic and transportation barriers to care (Lankowski, et al., 2014). Access to antiretroviral therapy
can be compared with patient outcomes, because patients who must travel far distances to receive services
may choose to forgo them. An assessment of fixed voluntary counseling and testing or voluntary medical
male circumcision (VMMC) may reveal the need for future scale-up or for mobile sites.
Finally, GIS can be used to integrate multiple pieces of data based on common geography. For instance,
it can be used to integrate information on patient load, available staff, and types of services offered at
facilities. This allows program managers to identify facilities that are under-performing or over-burdened
and nearby facilities that could serve as referral centers. Financial data, if available at disaggregate scales,
can be mapped against the services facilities offer to generate estimates of their efficiency. This process can
identify facilities that require support or supervision, and also where services need to be scaled back and
patients referred elsewhere. Investigating service coverage and outcomes in context with relevant ancillary
data including road corridors, employment opportunities, population density, and general demographics
can inform conversations about a health programs priorities.
Use of GIS to support the scale-up of HIV prevention and treatment activities in Tanzania
MEASURE Evaluation collaborated with USAID/Tanzania to build the capacity of local staff and
subcontractors to provide HIV programs in Iringa Region, in Tanzania. We worked with community
stakeholders to identify and map site locations of high-risk behaviors and the availability of HIV
prevention materials (e.g., condoms, flyers, etc.). We also mapped HIV prevention and care and
treatment sites and their catchment areas. When the maps showing locations of high-risk behaviors
were linked with catchment maps, program managers could identify facilities or community-based
organizations that could provide more outreach to those locations. The catchment maps were combined
with population data and service statistics to estimate service coverage provided for each site. By examining
unmet need, service area coverage, and priority populations together with factors such as travel distance
and physical barriers, program managers could identify gaps. For example, they could use the maps first
to identify an overburdened, high-volume site in a high-burden area, and then to identify alternative sites,
taking into consideration proximity as well as facility capacity. This allowed them to identify a nearby
clinic where care and treatment services could be scaled up.
Figure 1. HIV Voluntary Counseling and Testing Facility Coverage Compared with Population
Density, Iringa District, Tanzania
This map compares population density in Iringa Region, Tanzania with HIV voluntary counseling and testing facility
coverage to identify potential gaps in program access (Cunningham, et al., 2014).
The USAID-funded Maternal and Child Survival Program enhanced the geospatial data that MEASURE
Evaluation had collected with new information to guide male circumcision outreach and mobile services.
Additional health facility locations were collected. Areas of high HIV prevalence were linked with fixed
VMMC site locations and population data to identify priorities and generate estimates of the size of the
populations needing service. VMMC statistics from fixed and mobile services were linked to population
targets to allow program managers to track progress geographically. These maps were used to plan and
prioritize additional support and mentorship for facilities already offering services, to identify rural
facilities whose services could be scaled up, and to plan campaigns and mobile VMMC services based
on current need and historical performance. Information on road quality enabled program managers
to identify which facilities could be visited when; several were unreachable during the rainy season. The
result: Over a four-year period, Tanzanias Ministry of Health and Social Welfare was able to circumcise
more than 250,000 men ages 1035, raising the regional average from under 30 percent to over 80
percent.
This interactive map shows facility locations, VMMC program sites, and roads. This information enabled VMMC
program managers to identify low-performing facilities that might need support and determine whether they were
accessible in the rainy season or only in the dry season (Mahler, et al., 2015).
Using maps, program managers can build resilient health systems, by identifying and
prioritizing gaps in surveillance networks and services.
In addition to being a tool for directly mapping and analyzing epidemic trends, GIS can be used
to provide contexts useful in assessing risk, vulnerability, and resilience, by mapping areas at risk of
amplification and spread due to population mobility or structural factors (e.g., poverty and weak health
systems). Risk maps are typically developed based on historical outbreaks; on vector habitat; or on
locations of high human/animal interaction, where the animal hosts are known to carry zoonotic disease.
Vulnerability assessment maps may examine household size, housing materials, sanitation (latrines or
toilets), access to safe drinking water, or poverty. Together with population estimates, this information
can help programs designed to build resilience before an epidemic, provide context for response during
an epidemic, and support work to end an epidemic and rebuild health services. Rapid amplification and
spread are most likely to occur in situations where existing health services and surveillance are weak. Using
maps, program managers can build resilient health systems, by identifying and prioritizing gaps in existing
surveillance networks as well as in service provision due to lack of facilities, infrastructure, or trained staff.
Use of GIS to coordinate responses to the Ebola outbreak in West Africa
The World Health Organization (WHO), the Emergency Operations Center at the U.S. Centers for
Disease Control and Prevention, and the U.S. Global Development Lab and USAID GeoCenter used
GIS in the Ebola outbreak in Sierra Leone, Guinea, and Liberia. With it, they coordinated their responses,
identified areas at high risk, tracked new cases of the epidemic against prevalent areas while the epidemic
was in progress, and allocated resources.
GIS is also being used in reconstruction efforts to identify locations at high risk that could be prioritized
based on their social contexts. Reconstruction efforts include programs aimed at enhancing food security
through distribution and support to farmers; improving access to education for children while integrating
Ebola-related messaging in curricula as needed; and supporting the re-establishment of community- and
facility-based health services and enhancement of laboratory surveillance systems.
Figure 3. Conrmed and Probable Cases of Ebola in West Africa as of October 2014
New and recent cases of Ebola in West Africa as of October 29, 2014 are shown. This map, and others like it, allow
program managers to identify areas with current active transmission and previous transmission, and to quantify the
intensity of the epidemic in each area (WHO, 2014).
Figure 4. Map of Ebola Incidence in West Africa by Demographic and Household Variables
This web map provides necessary demographic context for program managers seeking to support Ebola
reconstruction efforts. It shows Ebola incidence, population density, average household size, the proportion of
female-headed households, access to toilets or latrines, and access to safe drinking water (USAID, n.d.).
Using Geographic Information Systems to Meet Global Health Challenges
This map (and others like it) shows the cumulative number of Ebola cases, the subnational areas where new cases
are occurring, and which partners, funded by which U.S. government agency, are involved in what services in
which countries. It provides a high-level view of who is doing what and where to improve coordination (Corbett,
2015).
GIS has been used extensively for assessing and improving access to basic
emergency obstetric care (BEmOC) and comprehensive emergency obstetric care
(CEmOC).
Global recommendations for access to CEmOC state that access within a two-hour travel period is
critically important for the mothers survival if she has complications during delivery. In scaling up
CEmOC services to meet this need, GIS has been used to map existing BEmOC and CEmOC services,
travel times to these services, and population density. By correlating these three factors, program managers
can identify areas where large populations lack services and identify BEmOC sites near those populations
that could be upgraded to CEmOC. This was done in Uganda, where upgrading facilities improved
physical access to services.
Figure 6. Access to Comprehensive Emergency Obstetrical Care in Western Uganda Before
and After Interventions to Improve It
This map shows improvement in CEmOC service coverage in western Uganda, because of an intervention by
Saving Mothers, Giving Life. The map on the left depicts the distribution of facilities providing CEmONC service
before the intervention; the map on the right shows improved access after the intervention was implemented
(Serbanescu, 2014).
This map presents satellite imagery together with GPS points showing areas where a polio immunization team has
visited in northern Nigeria. When aggregated to a larger scale and linked with village geo-coordinates, maps like
this can show at a glance which villages or households were missed and need to be visited during campaign mopup (Gates, 2012).
Figure 8. Material Used by Polio Immunization Teams to Plan Campaigns in Northern
Nigeria
Printed maps and atlases of local government areas and towns in northern Nigeria were used by polio immunization
teams to improve plans for campaigns and to ensure that every village was reached (Gates, 2012).
10
GIS maps can be viewed at multiple scales (facility, district, region) to meet the needs
of multiple audiences.
Three overlapping factors make geospatial tools uniquely suited to accomplishing these four tasks.
First, GIS can frame data within common geographic contexts to allow program managers to layer or
integrate health and ancillary information from different sources, considering data from each source
alone as well as in relation to data from the other sources. Linking data in this way can provide necessary
perspective to inform discussions on resource allocation or program strategies. For example, program
managers can align human resources data with the number of patients visiting each facility, to identify
facilities that are overburdened or have extra capacity. Public health data can be linked to data on ecology
for an assessment of malaria risk; to socioeconomic data for an analysis of social vulnerability; or to a
combination of data on population, land cover (the physical material on a geographic areas surface), roads,
and service delivery points to examine gaps in program access.
Figure 9. Prevalence of Malaria among Children Ages 210 in Africa
The Malaria Atlas Project combines information on malaria prevalence from surveys, environmental factors, and
demographic factors to estimate malaria (Plasmodium falciparum) infection prevalence rates in children ages 210
(Bhatt, et al., 2015).
11
Second, GIS gives program managers a geographic perspective on health statistics for a given location in
relation to adjoining areas. If an area with low prevalence of an infectious disease is adjacent to one with
high prevalence, it may be at higher risk for its disease burden to increase, because the disease might be
transmitted along paths or roads from the neighboring area. Clusters of high prevalence might indicate
underlying systemic vulnerabilities that need to addressed.
Third, GIS gives program managers an information product that can spur discussions. Maps are powerful
visual images providing a shared context for discussions. They can be viewed at multiple scales (facility,
district, region) to meet the needs of multiple audiences.
Better data and better tools have increased the capacity to use GIS worldwide. GIS
software can be installed and staff can be trained easily.
Recent years have seen burgeoning cadres of M&E staff and data science staff with basic GIS skills able to
create maps or perform simple geographic analyses. This means that an organization does not necessarily
need dedicated GIS staff. The growth in the number of professionals having GIS skills can be attributed
partly to improvements in software availability as well as usability; increasing demand by decision
makers for geospatial information products; and increased availability of training resources and guidance
12
documents designed for non-GIS professionals. Such resources include both big-picture guidance and
step-by-step instructions for key GIS tasks. The use of GIS in data science has grown in step with the
data science field, and also as part of a general movement toward assembling and analyzing large data sets.
Further, as databases such as DHIS 2 collect vast amounts of health data, the ability to routinely process
and convert high volumes of data into meaningful information has become more important.
The abilities to process, investigate, and use high volumes of data at multiple scales and to set up analyses
that can be conducted routinely with only marginal effort allow managers to adjust their programs
frequently throughout the year.
Data
Two recent trends in geographically oriented health data are:
Increased density of information at high resolutions
Increased availability of contextual layers, also at high resolution
Improvements in technology are driving these trends. Some examples: Tools such as smart phones have
simplified the collection of geospatial data. Pre-processed satellite imagery is readily available. As GIS
technology and software gain traction, demand for spatial data has increased. As the development field
shifts to open data, the importance of local context is becoming clearer, as is the importance of identifying
and addressing pockets or gaps in health coverage.
Contextual data and the interplay among multiple layers of datafor example, high resolution
disaggregated population data, satellite imagery, and maps of road networks, rivers, and terrainprovide
new perspectives on the determinants of a programs success. Routine health information systems and
IDSRs are becoming more sophisticated, and often provide health data at high resolutionat the facility
or even community level. Techniques for taking this data, as well as data from surveys, and creating
smoothed maps is increasing access to risk-maps by program staff, and with it, the ability of staff to set
effective and efficient priorities for their interventions.
As routine health information systems improve, access to high-quality data at the facility level has become
the new norm. In this data environment, integration of GIS with data is becoming easier. If the facilities
are geographically located (have geo-codes), linking a routine health information system (RHIS) with
a GIS or a decision support system (as has been done within the DHIS 2) allows maps to be generated
routinely as new information becomes available. Although setting up the initial link between the GIS and
RHIS requires substantial initial effort, the pay-off can be significant, as program managers and M&E
officers are able to visualize their data rapidly.
Conclusion
As countries make progress toward global development goals, challenges and program gaps will become
increasingly local. Programs will go from a focus on reaching every district to reaching every community,
and increasingly they will need access to subnational and even subdistrict data. Geospatial toolswith
their ability to examine and overlay multiple factors at once; to examine health information across
multiple scales; and to provide geographic contextare uniquely suited to support health programs
13
in this environment. Maps and other GIS-produced information products can illuminate spatial and
temporal trends in communities, support program planning, and enhance advocacy messages. Barriers to
GIS use continue to fall. GIS is increasingly integrated in or linked with existing data collection or data
management systems. Improvements in routine health information systems have provided a wealth of
geographically disaggregated data.
The geospatial environment provides program managers with an opportunity to leverage the power
of GIS to improve program efficiency as well as address equity gaps.
14
References
Bhatt, S., Weiss, D. J., Cameron, E., Bisanzio, D., and Mappin, B., et al. 2015. Th e E ff ect of Malaria
Control on Plasmodium falciparum in Africa between 2000 and 2015. Nature 526: 207-211. Available at:
http://www.nature.com/nature/journal/vaop/ncurrent/full/nature15535.html.
Chabot-Couturea, G., Seaman,V.Y., Wenger, J.,Moonen, B.,and Magill, A. 2015. International
Health 7(2): 7981. Advancing Digital Methods in the Fight against Communicable Diseases.
http://inthealth.oxfordjournals.org/content/7/2/79.abstract.
Corbett, N. From Vaccinations to Vitamins: Ensuring West Africans Get Critical Care Amid Ebola
Crisis. U.S. Agency for International Development (USAID) Impact, April 18, 2015. Available at:
https://blog.usaid.gov/2015/04/from-vaccinations-to-vitamins-ensuring-west-africans-get-critical-careamid-ebola-crisis/.
Cunningham, M., Mapala, Y., Patrick, J. 2014. Using Geospatial Analysis to Improve Resource
Allocation for HIV Programs in Iringa Region, Tanzania. Available at: http://www.cpc.unc.edu/measure/
publications/sr-14-107.
Gates, B. Digital Mapping Technology Helps Polio Vaccinators Zero In. Gates Notes, November 5, 2012
http://www.gatesnotes.com/Health/Digital-Mapping-Technology-Helps-Polio-Vaccinators-Zero-In.
Global Health Sciences, University of California, San Francisco. n.d. Google Fusion Tables for Health.
Available at: https://datause.ucsf.edu/google-fusion-tables-health.
Lankowski, A.J.,Siedner, M.J.,Bangsberg, D.R.,and Tsai, A.C. 2014. Impact of Geographic and
Transportation-Related Barriers on HIV Outcomes in Sub-Saharan Africa. AIDS and Behavior
18(7):1199-223. Available at: http://www.ncbi.nlm.nih.gov/pubmed/24563115.
Lee, S.S. and Wong, N.S. 2011. The Clustering and Transmission Dynamics of Pandemic Influenza A
(H1N1) 2009 Cases in Hong Kong. Journal of Infection 63(4):274-80. Available at:
http://www.ncbi. nlm.nih.gov/pubmed/21601284.
Mahler, H., Searle, S., Plotkin, M., Kulindwa, Y., and Greenberg, S., et al. 2015. Covering the Last
Kilometer: Using GIS to Scale-Up Voluntary Medical Male Circumcision Services in Iringa and Njombe
Regions, Tanzania. 2015. Global Health Science and Practice 3(3): 503515. Available at: http://www.
ghspjournal.org/content/3/3/503.full.pdf.
MEASURE Evaluation. 2013. Geographic Tools for Global Public Health: An Assessment of Available
Software. Available at: http://www.cpc.unc.edu/measure/publications/ms-13-80.
MEASURE Evaluation. 2014. March 2014 MEASURE GIS Working Group Meeting: Th e E volution
of GIS for Planning, Monitoring and Evaluation in Global Health over the Past Five Years. Available at:
http://www.cpc.unc.edu/measure/our-work/gis/measure-gis-working-group-meeting/measure-gisworking-group-march-2014-meeting-summary.
15
Serbanescu, F. Saving Mothers Giving Life Initiative: Experiences with Mapping Maternal Deaths
Uganda. Paper presented at Maternal Mortality Mapping Workshop, Washington DC, 2014.
Tanser, F.,Brnighausen, T.,Cooke, G.S.,and Newell, M.L. 2009. Localized Spatial Clustering of HIV
Infections in a Widely Disseminated Rural South African Epidemic. International Journal of Epidemiology
38(4):1008-16. Available at: http://www.ncbi.nlm.nih.gov/pubmed/19261659.
USAID. n.d. Demographic Data for Ebola Response. Available at:
http://usaid.maps.arcgis.com/apps/MapSeries/?appid=c67114891e0f4ca5ba74269db64e23bd.
Weir,S.S., Morroni, C., Coetzee, N., Spencer, J., and Boerma, J.T. 2002. Pilot Study of a Rapid
Assessment Method to Identify Places for AIDS Prevention in Cape Town, South Africa. Sexually
Transmitted Infections 78:i106-i113. Available at: http://sti.bmj.com/content/78/suppl_1/i106.
World Health Organization (WHO). 2014. Ebola Outbreak Response: Maps, 2014. Available at: http://
www.who.int/csr/disease/ebola/maps/en/.
Zulu LC,Kalipeni E,and Johannes E. 2014. Analyzing Spatial Clustering and the Spatiotemporal Nature
and Trends of HIV/AIDS Prevalence Using GIS: The Case of Malawi, 1994-2010. BMC Infectious
Diseases 14:285. Available at: http://www.ncbi.nlm.nih.gov/pubmed/24886573.
16
17
MEASURE Evaluation
University of North Carolina at Chapel Hill
Carolina Population Center
400 Meadowmont Village Circle, 3rd Floor
Chapel Hill, NC 27517
http://www.measureevaluation.org